An Optimization Framework for Data Collection in Software Defined Vehicular Networks.

Area of Science:

• Computer Science
• Electrical Engineering
• Network Engineering

Background:

• Software Defined Vehicular Networks (SDVN) offer enhanced programmability in Vehicular Adhoc Networks (VANETs).
• Existing SDVN architectures lack efficient data collection mechanisms.
• Dynamic network topologies necessitate adaptive optimization strategies.

Purpose of the Study:

• To propose a novel data collection methodology for hybrid SDVN architectures.
• To optimize data collection by minimizing agents, delay, cost, payload, and overhead.
• To address the challenge of dynamic network topologies in vehicular communications.

Main Methods:

• Modeling data collection as an Integer Quadratic Programming (IQP) problem.
• Optimally selecting broadcasting and agent nodes.
• Utilizing heterogeneous normalized network link entropy for systematic data collection and optimization decisions.

Main Results:

• The proposed IQP model significantly reduces communication cost by 75.5% and end-to-end latency by 32.7% compared to centralized SDVN architectures.
• Achieved 99.9% data collection rate in large-scale vehicular networks under optimized settings.
• Demonstrated the effectiveness of inspecting network link entropy for dynamic optimization.

Conclusions:

• The novel data collection methodology provides an efficient and optimized solution for hybrid SDVN architectures.
• The IQP model effectively balances multiple objectives for improved vehicular network performance.
• This approach enhances the flexibility and efficiency of data management in dynamic vehicular environments.